Coupling WOFOST and HYDRUS-1D to simulate daily maize growth, water– salt dynamics and stress responses under salinity gradients in salinized farmland

Soil salinization constrains agricultural productivity across approximately 950 million hectares worldwide. In the Yellow River irrigation district of Ningxia, secondary salinization severely depresses maize yields. Existing crop–water–salt models lack day-by-day bidirectional coupling between salt transport and crop growth, over-simplify salt stress representation, and amplify stress through multiplicative integration. To address these gaps, we developed a fully coupled WOFOST–HYDRUS-1D model linking the Richards equation and convection–dispersion equation with crop photosynthesis, transpiration, and assimilate partitioning through a modified Maas–Hoffman function. Salt stress is transmitted via three physiological pathways, and the combined stress factor is computed using Liebig’s law of the minimum. The model was calibrated with 43 sampling points spanning low-to-high salinity gradients (1.49–6.79 g kg⁻¹) in Huinong District during 2024, and independently validated with 30 points (1.29–7.75 g kg⁻¹) in 2025. Calibration yielded R² = 0.883, RMSE = 0.744 t ha⁻¹, NRMSE = 10.96%, and NSE = 0.795; validation gave R² = 0.824, RMSE = 0.753 t ha⁻¹, NRMSE = 11.13%, and NSE = 0.810, confirming strong inter-annual parameter stability. Under high salinity (> 4 g kg⁻¹), simulated mean yield declined to 3.75 t ha⁻¹, a 53% reduction compared with low-salinity conditions. Compared with the standard WOFOST model (R² = 0.450, RMSE = 1.399 t ha⁻¹), the coupled model substantially improved accuracy.These results show that the coupled model can improve yield prediction under salinity stress and provide a useful tool for irrigation scheduling and water–salt management in salinized farmland.